Robustness of Rotation-Equivariant Networks to Adversarial Perturbations

Beranger Dumont, Simona Maggio, Pablo Montalvo
2/19/2018
cs.CVcs.CRcs.LG

Official code: rakutentech/stAdv (75 stars, MIT, TensorFlow, last commit 2018-11-03).

Abstract

Deep neural networks have been shown to be vulnerable to adversarial examples: very small perturbations of the input having a dramatic impact on the predictions. A wealth of adversarial attacks and distance metrics to quantify the similarity between natural and adversarial images have been proposed, recently enlarging the scope of adversarial examples with geometric transformations beyond pixel-wise attacks. In this context, we investigate the robustness to adversarial attacks of new Convolutional Neural Network architectures providing equivariance to rotations. We found that rotation-equivariant networks are significantly less vulnerable to geometric-based attacks than regular networks on the MNIST, CIFAR-10, and ImageNet datasets.

View on arXivView PDF

Code Implementations(1)

rakutentech/stAdvOfficial100%
TensorFlowMIT

Spatially Transformed Adversarial Examples with TensorFlow

7518Python, MakefileMay 11, 20187 years agoMIT
adversarial-exampletensorflow

Similar Papers

Exploring the Landscape of Spatial Robustness

88% match2 repos472017

Not So Robust After All: Evaluating the Robustness of Deep Neural Networks to Unseen Adversarial Attacks

85% match2023

Adversarial Attacks on Hyperbolic Networks

85% match2024

Bridging Symmetry and Robustness: On the Role of Equivariance in Enhancing Adversarial Robustness

84% match2025

Fooling Neural Networks for Motion Forecasting via Adversarial Attacks

84% match2024

Boosting the Transferability of Adversarial Attacks with Reverse Adversarial Perturbation

84% match2 repos352022

Cite this paper

@article{dumont2018robustness,
  title  = {Robustness of Rotation-Equivariant Networks to Adversarial Perturbations},
  author = {Beranger Dumont and Simona Maggio and Pablo Montalvo},
  year   = {2018},
  eprint = {1802.06627},
  archivePrefix = {arXiv},
  url    = {https://arxiv.org/abs/1802.06627}
}

Add a badge to your README

Shows whether official code is linked for this paper.

Code of Paper badge
[![Code of Paper](https://codeofpaper.com/badge/1802.06627.svg)](https://codeofpaper.com/paper/1802.06627)

Discussion